Re: UDB question

["R. Bruce Hoffman, Jr." <rbruceh@xxxxxxxxxxxxx>]

----- Original Message -----
From: "Dan Rasch" <drasch@mail.win.org>
To: <midrange-l@midrange.com>
Sent: Wednesday, December 19, 2001 1:37 PM
Subject: UDB question


> I was just told there is a directive from management to look into
> UDB for the 400's DB2.  Apparently the claim is that if we replace
> our DDS with SQL definitions, we can change file defs (like adding
> fields to the end of the record) without the level checks.

Level checks are not caused by the definition (DDL of SQL) but the access
and manipulation of data (DAL of SQL) through programs that refer directly
to the file and it's external definition.

SQL _access_ (select, delete, insert, update) if done _without_  the * (ie.
select * from a into bstruct) will ignore level differences on the record
because it looks at the field level, not the record level for its data.

As for UDB, on the 400 it's just a name. If you have an AS400 or iSeries,
then you have "DB2 UDB for OS/400". There are three flavors (and producers)
of DB2 UDB. One runs on the 400, one runs on 390 architecture and (IMO the
true UDB) the last runs on just about everything else with a common code
base and comes from Toronto.

The 400's version of DB2, in many cases, lags behind the facilities of the
"all platform" version of UDB and the 390's version. The glaring exception
was EVIs which made it to the 400 first. But the 400 still does not support
grouping sets, rollup and cube operations, although it finally has
declaritive triggers, and more than 6 allowed.

Also IMO, switching the definition to SQL allows you to use tools for
database design, modelling, creation, maintenance and documentation. All
well worth the effort.

===========================================================
R. Bruce Hoffman, Jr.
 -- IBM Certified Specialist - iSeries Administrator
 -- IBM Certified Specialist - RPG IV Developer

"I want to be different, just like everybody else!"
  - Ceili Rain